Trajectory Tracking and Obstacle Avoidance of Robotic Fish Based on Nonlinear Model Predictive Control.

Area of Science:

• Robotics
• Control Systems
• Marine Engineering

Background:

• Accurate motion control for robotic fish in complex underwater environments remains a significant challenge.
• Existing methods often struggle with real-time adaptation and precise maneuverability.

Purpose of the Study:

• To propose and evaluate a nonlinear model predictive control (NMPC) algorithm for enhanced trajectory tracking and obstacle avoidance in robotic fish.
• To demonstrate the real-time implementation and effectiveness of NMPC for agile underwater navigation.

Main Methods:

• Formulation of a dynamic model for the robotic fish, including accelerations, within a world coordinate system.
• Development of NMPC with obstacle avoidance and objective functions, incorporating control restrictions.
• Comparative analysis of NMPC against Pure Pursuit (PP) for trajectory tracking and Dynamic Window Approach (DWA) for obstacle avoidance.

Main Results:

• NMPC demonstrated superior accuracy in trajectory tracking compared to the Pure Pursuit algorithm.
• NMPC exhibited greater resilience in obstacle avoidance planning than the Dynamic Window Approach.
• The proposed NMPC controller effectively manages robotic fish movements with high agility.

Conclusions:

• The NMPC-based strategy provides a viable solution for precise trajectory tracking and efficient obstacle avoidance in robotic fish.
• This approach shows significant potential for practical application in complex underwater robotics scenarios.